Chronic hepatitis C virus (HCV) infection affects

Transcription

1 GASTROENTEROLOGY 2002;123: Adherence to Combination Therapy Enhances Sustained Response in Genotype-1 Infected Patients With Chronic Hepatitis C JOHN G. MCHUTCHISON,* MICHAEL MANNS, KEYUR PATEL,* THIERRY POYNARD, KAREN L. LINDSAY, CHRISTIAN TREPO, JULES DIENSTAG, # WILLIAM M. LEE,** CARMEN MAK, JEAN JACQUES GARAUD, andjanice K. ALBRECHT, FOR THE INTERNATIONAL HEPATITIS INTERVENTIONAL THERAPY GROUP *Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina; Division of Gastroenterology andhepatology, Medical School of Hannover, Hannover, Germany; Groupe Hospitalier Pitié-Salpétrière, Paris, France; Division of Gastroenterology and Hepatology, University of Southern California, Los Angeles, California; Service d Hépato-Gastroénterologie, Hotel Dieu, Lyon, France; # Massachusetts General Hospital, Gastrointestinal Unit, HarvardMedical School, Boston, Massachusetts; **University of Texas Southwestern Medical Center, Dallas, Texas; and Schering-Plough Research Institute, Kenilworth, New Jersey The investigators andinstitutions participating in the Hepatitis Interventional Therapy Group are listedin the appendix. Background & Aims: Patient adherence to prescribed antiviral therapy in human immunodeficiency virus infection enhances response. We evaluated the impact of adherence to combination therapy with interferon or peginterferon plus ribavirin in chronic hepatitis C patients. Methods: We assessed the effect of dose reduction on sustained virologic response (SVR) from prior trials with interferon -2b plus ribavirin (n 1010) or peginterferon -2b 1.5 g/kg/week plus ribavirin (n 511). The actual treatment administered was verified from drug dispensing/return records and patient diaries. Two groups were defined: (1) patients who received >80% of both their total interferon and ribavirin doses for >80% of the expected duration of therapy and (2) patients who received reduced doses (<80% of one or both drugs for >80% of the expected duration of therapy). A statistical model provided comparative estimates of the response rates in compliant patients. Results: Most patients were at least 80% compliant with interferon -2b/ribavirin or peginterferon -2b/ribavirin therapy and had SVR rates of 52% and 63%, respectively, for the 2 regimens. This was most apparent for HCV-1 infected patients. The impacts of adherence on efficacy from subgroup analysis and the statistical modeling approach were similar. Conclusions: HCV-1 infected patients who can be maintained on >80% of their interferon or peginterferon -2b and ribavirin dosage for the duration of treatment in the setting of a clinical trial exhibit enhanced sustained response rates. Our results suggest that adherence will enhance the likelihood of achieving an initial virologic response. Adherence beyond weeks will be advantageous only for those patients who have achieved such an early virologic response. Chronic hepatitis C virus (HCV) infection affects approximately 300 million people worldwide and is the most common cause of chronic liver disease and the most frequent indication for liver transplantation in the United States. 1,2 The most effective currently available therapy is the combination of -2b interferons (pegylated or nonpegylated) plus ribavirin. Between 41% and 47% of patients treated with interferon -2b plus ribavirin achieve a sustained virologic response, which seems to be durable, long-lasting, and associated with potential long-term benefits. 3 7 More recently, the combination of peginterferon -2b plus ribavirin has been shown to improve this response rate to 54%, 7 with a secondary retrospective analysis indicating a response rate of 61% when the doses of both peginterferon -2b and ribavirin are evaluated in terms of patient weight. A prospective trial evaluating weight-based dosing of ribavirin combined with peginterferon -2b is currently in progress to further evaluate this concept. Therapy with a combination of interferon -2b or peginterferon -2b and ribavirin requires a moderately complex regimen of subcutaneous injections, twice-daily oral administration, frequent visits with blood tests to monitor safety, and side effects in nearly all patients. As a result, not all patients complete their course of treatment. In all likelihood, patient benefit and response Abbreviations used in this paper: HCV, hepatitis C virus; HIV, human immunodeficiency virus; ITT, intention to treat; SVR, sustained virologic response by the American Gastroenterological Association /02/$35.00 doi: /gast

2 1062 McHUTCHISON ET AL. GASTROENTEROLOGY Vol. 123, No. 4 Figure 1. Schematic diagram of the 4 studies performed with regimens of interferon -2b or peginterferon -2b plus ribavirin, 3,4,7 and peginterferon -2b or interferon -2b alone. 10 Asterisks indicate the groups of patients where adherence was evaluated from pharmacy records, pill counts, andpatient diaries. For patients receiving interferon -2b plus ribavirin, we evaluatedadherence in all patients receiving 48 weeks of therapy, those with HCV-1 infection receiving 48 weeks of treatment, andthose with HCV-2 or -3 infection receiving either 24 or 48 weeks of treatment. From the Manns et al. study, 7 we also evaluatedadherence in those receiving the most effective regimen, 48 weeks of peginterferon -2b 1.5 g/kg/week plus ribavirin for 48 weeks first, in all patients in the subgroups with HCV-1 or HCV-2 or -3, andalso in the subgroup receiving the weight-baseddosing regimen of ribavirin 10.6 g/kg, as shown. From the Lindsay et al. study, 10 we evaluatedthe peginterferon -2b 1.5 g/kg andinterferon -2b monotherapy groups, once again in all patients, andalso according to genotype. In each subgroup, the number of patients that couldbe evaluatedis noted. would be improved with better adherence to the prescribed treatment regimen. In human immunodeficiency virus (HIV) infection, current treatment regimens with highly active antiretroviral therapy (HAART) are also complex and associated with side effects. Nonetheless, strict adherence to the prescribed therapeutic regimen has been shown to improve the virologic response. 8,9 Based on the aforementioned issues and the similarities between HCV and HIV infections, we undertook this study. Our primary aim was to evaluate the effect on sustained response rate of adherence to combination therapy (interferon -2b or peginterferon -2b plus ribavirin) in patients with chronic hepatitis C. Materials and Methods Patient Selection Data for patients from 3 previously published clinical trials of interferon -2b or peginterferon -2b and ribavirin combination therapy were evaluated retrospectively to determine the effect of dose reduction on sustained response (Figure 1). 3,4,7 In all, 1010 patients treated with interferon -2b (INTRON A; Schering-Plough, Kenilworth, NJ) and ribavirin (REBETOL; Schering-Plough) and 511 patients treated with peginterferon -2b (PEG-Intron; Schering-Plough) plus ribavirin were considered in this analysis. These subgroups of patients evaluated in the aforementioned studies had been randomized to receive either interferon -2b 3 times a week (for 24 or 48 weeks) in combination with daily ribavirin or, alternatively, peginterferon -2b at a dose of 1.5 g/kg per week subcutaneously as a single injection with ribavirin 800 mg daily for 48 weeks. These patient subgroups were selected because they were those that received the most effective regimens in these respective trials. Data from a separate monotherapy trial 10 of 304 patients treated with peginterferon -2b, 1.5 g/kg per week, and 303 patients who received interferon -2b 3 times a week for 48 weeks were also analyzed to evaluate the effect of adherence on response in this trial. All patients had given informed consent for the study in which they had participated, and the ethics committee at each

3 October 2002 ADHERENCE TO THERAPY AND HEPATITIS C 1063 clinical site had approved each study. The databases for these studies were created and maintained centrally and included age, sex, ethnic background, presumed duration of infection, body weight, and mode of infection. Histologic activity was assessed by a single pathologist blinded to the treatment regimen and timing of the liver biopsies and graded by the Histologic Activity Index. 11 Serum alanine aminotransferase (ALT) and HCV RNA values were measured before, during, and at the end of therapy and then for 6 months after treatment. The HCV genotype was also recorded for each patient. Sustained virologic response was defined as the absence of detectable serum HCV RNA by reverse-transcription polymerase chain reaction 24 weeks after the completion of therapy. Serum HCV RNA testing was performed at a single central laboratory (National Genetics Institute, Los Angeles, CA) by a quantitative polymerase chain reaction with a detection limit of 100 copies/ml. 12 HCV genotyping was performed with INNOLIPA HCV (Innogenetics, Zwijnaarde, Belgium) as described previously. 13 We took 2 approaches to assessing the effect of adherence on sustained virologic response. The first approach assigned patients who received combination therapy into subgroups according to their adherence. The second approach incorporated adherence as a covariate in a statistical model. Subgroup Analysis For this analysis, the amount of each drug administered to a patient was obtained from drug dispensing/return records and patient dosing diaries. Patients receiving combination therapy were divided into 2 groups for analysis: (1) the 80/80/80 subgroup, comprising patients who were 80% adherent (i.e., received 80% of their total interferon dose and 80% of the ribavirin dose) and were treated for 80% of the expected duration of therapy; and (2) the 80/ 80/80 subgroup, comprising patients who underwent dose reduction ( 80% of 1or both drugs for 80% expected duration). Patients who withdrew from the study prematurely were excluded from the analysis. The goal of 80% of the planned drug dosage for 80% of the assigned duration represents an adherence criterion that had been adopted previously in the assessment of the efficacy of other pharmaceutical agents, including HIV drug therapy, antihypertensive drug therapy, and orally administered cancer drug therapies. 8 Sustained response rates were then determined for all patients in each of the subgroups receiving interferon -2b3MU three times a week plus ribavirin mg/day, peginterferon -2b 1.5 g/kg every week plus ribavirin 800 mg/ day, and the patients who received therapy with peginterferon -2b 1.5 g/kg every week plus weight-based dosing with ribavirin of 10.6 mg kg 1 day 1. The analysis for the same treatment groups was performed based on genotype. The analysis of these subgroups reflects observed differences in patients who are most adherent compared with those who are not. On a post hoc retrospective analysis of this type, however, in the absence of a control variable for adherence, the potential exists for adherence to be affected by selection bias. Statistical Modeling To overcome the possibility of overestimating the treatment effect due to bias, we relied on a statistical framework developed by Efron and Feldman 14 that had been adapted to data from hepatitis C clinical trials by Mak et al. 15 This methodology allows the objective incorporation of retrospective adherence data in a statistical model to estimate the true efficacy response for the intended treatment at full adherence, and also serves as a reference point for the subgroup analysis. This analysis relies on several key assumptions. First, the model assumes a monotone-increasing relationship crossing the origin between medication dose and efficacy response (i.e., the dose-response curve). Second, whereas a true dose-response curve is unlikely to be observed in any clinical therapeutic trial due to possible nonadherence, in our model dose-response is assumed to be related to the observed adherence-response pattern. Further, the model assumes that adherence is driven by a patient s underlying propensity to adhere and is not dependent on other covariates, either observed or unobserved. Finally, it is assumed that the adherence propensity can be modeled by adherence-response data from a compatible control in which patients derived minimal efficacy directly due to the control medication. Under these assumptions, an estimate of the sustained virologic response at full dose is possible. The steps of the analysis are briefly summarized here. The goal of the analysis was to estimate the value of the true efficacy response of the intended combination therapy from the observed adherence-response relationship. To estimate the adherence-response relationship, we fitted a logistic regression to the combination therapy data and to its corresponding control monotherapy data. The arithmetic difference between the 2 adherence-response regression curves was used to estimate the dose-response curve of efficacy of combination therapy above that of control monotherapy at full adherence (100%). Finally, the intention-to-treat (ITT) efficacy rate of control monotherapy was added back to this estimate to yield the fulladherence efficacy of combination therapy. Six models were fitted. The first model was fitted to all patients who received interferon -2b plus ribavirin and all patients who received interferon -2b plus placebo; the second model, to all patients who received peginterferon -2b and ribavirin or peginterferon -2b alone. The next 2 models were similar to the first 2 but limited to patients with HCV-1. Finally, the last 2 models were fitted to all patients and HCV-1 infected patients separately who received peginterferon -2b and weight-based dosing of ribavirin ( 10.6 mg kg 1 day 1 ); thus, both subsets included only patients weighing 75.5 kg (ratio of 800 mg/day of ribavirin to 10.6 mg kg 1 day 1 ). The control groups for the last 2 models consisted of all patients and HCV-1 infected patients who weighed 75.5 kg and who received monotherapy. From each of the models, we estimated the sustained virologic response rate for combination therapy at full dose. For the purpose of this statistical model, we included an additional 503 patients treated with interferon -2b plus

4 1064 McHUTCHISON ET AL. GASTROENTEROLOGY Vol. 123, No. 4 placebo and 304 patients treated with peginterferon -2b 1.5 g/kg alone from these studies. 3,4,7,10 Each of the 2 treatment groups receiving interferon -2b or peginterferon -2b monotherapy was used as a control group for its respective interferon ribavirin combination-therapy groups. Statistical Analysis Differences in baseline characteristics between the groups were summarized descriptively and were assessed by 2 analysis and the 2-sided t test. Response rates were compared by 2 analysis. Results Of the 1010 patients evaluated who received interferon -2b plus ribavirin, 218 were excluded from further analysis because the duration of therapy was 80% of the assigned treatment regimen. Most of the remaining patients who were treated for longer than 80% of the planned duration, 631of 792 (80%), also received 80% of both their interferon -2b and ribavirin doses. Similarly, for the 511 patients receiving peginterferon -2b 1.5 g kg 1 wk 1 plus ribavirin, 88 were excluded from the analysis as indicated earlier, and of the remaining 423, 305 (72%) were adherent and received 80% of both medications as defined previously. Of the 304 patients who received peginterferon -2b 1.5 g kg 1 wk 1 alone, 45 (15%) were excluded because they failed to achieve at least 80% of the assigned duration of therapy. Likewise, 56 (18%) individuals were excluded from the 303 patients who received interferon -2b alone. Most of the remaining patients in these monotherapy trials were adherent to therapy, with 236 of 259 (91%) and 242 of 247 (98%) receiving 80% of peginterferon -2b 1.5 g kg 1 wk 1 and interferon -2b alone, respectively. Overall, patients who received 80% of 1or both drugs for 80% of the duration of therapy had lower sustained response rates than the subgroups that received 80% of the assigned duration of therapy: peginterferon -2b and ribavirin, 25% (22 of 88); interferon -2b plus ribavirin, 18% (39 of 218); interferon -2b monotherapy, 0% (0 of 56); and peginterferon -2b monotherapy, 4% (2 of 45). Inclusion of these patients in subsequent analyses would have biased the results favorably in terms of the effect of adherence on response; hence the rationale for excluding this group of patients from subsequent analysis in this study. The main reasons for not achieving the adherence goals outlined earlier were adverse events to therapy in 75% of the patients. In the remaining patients, failure to attend scheduled appointments, withdrawal of consent, and nonadherence in the absence of apparent side effects were the other reasons for not achieving adherence targets in this study. When evaluating the patients who were most adherent to combination therapy compared to the others, we found that baseline viral and host characteristics of the 2 groups were similar in terms of HCV genotype, viral load, and ALT values (Table 1). A larger proportion of adherent patients was male and weighed more. The patients who received interferon -2b and ribavirin but were less adherent were older (P 0.047) and had a higher incidence of advanced fibrosis (34% vs. 24%; P 0.015) than adherent patients receiving the same therapy. These differences were not apparent in the adherent and less-adherent subgroups receiving peginterferon -2b and ribavirin. For the subgroup analysis, adherence affected the sustained response rates for all of the regimens that we examined. HCV-1 infected patients who were most compliant with therapy had significantly higher sustained virologic response rates (Table 2). This observation was evident for patients receiving interferon -2b plus ribavirin or peginterferon -2b plus ribavirin for 48 weeks, and was not apparent for those treated with peginterferon -2b or interferon -2b monotherapy. Adherence did not appear to enhance virologic response in patients with HCV-2 or -3 infection treated for either 24 or 48 weeks. Sustained virologic response rate as a function of the amount of treatment received was assessed at different adherence levels to determine any relationship between adherence and sustained response. As shown in the data for Figures 2, 3, and 4, we found a continuous, increasing relationship between adherence and sustained virologic response. Thus at least 80% adherence to therapy increased sustained virologic response rates for patients who received interferon -2b plus ribavirin from 44% to 52%, for patients who received peginterferon -2b plus ribavirin from 54% to 63%, and for patients who received peginterferon -2b plus weight-based ribavirin from 61% to 72%. The same adherence-response relationships were observed in patients with HCV-1and in patients who received weight-based dosing of peginterferon -2b plus ribavirin. In contrast, poorer adherence to therapy of 20% reduced overall sustained response rates to only 16% to 17% in the treatment arms. To supplement the interpretation of these data, and to overcome potential bias introduced by the subgroup analysis, we provide the results of the analysis based on statistical methodology that takes into account the retrospective nature of the adherence data. Figure 2 shows the raw sustained response rates observed at different

5 October 2002 ADHERENCE TO THERAPY AND HEPATITIS C 1065 Table 1. Pretreatment Characteristics of Treatment Groups According to Adherence IFN -2b Ribavirin PEG IFN -2b Ribavirin P value 80/80/80 a 80/ 80/ 80 b 80/80/80 c 80/ 80/ 80 d a vs. b c vs. d N M/F 452/179 75/86 204/101 55/ Age 35 21% 17% 15% 12% % 71% 78% 79% NS 56 6% 12% 7% 8% Weight 65 kg 16% 32% 12% 25% kg 46% 43% 48% 34% kg 38% 25% 40% 41% Source of infection Transfusion 20% 31% 22% 25% Parenteral 56% 50% 59% 68% Other 24% 19% 19% 7% Genotype 1 67% 67% 68% 67% NS NS 2 or 3 30% 30% 29% 32% HCV RNA 2M 68% 65% 67% 67% NS NS ALT X ULN e (mean SD) NS NS Bridging fibrosis or cirrhosis 24% 34% 27% 36% NS NS, Not significant. a Denotes the subgroup receiving 80% interferon -2b plus 80% ribavirin for more than 80% of the expectedduration of therapy. b Denotes the subgroup receiving 80% interferon -2b and/or 80% ribavirin for 80% expectedduration of therapy. c Represents the subgroup receiving 80% peginterferon -2b plus 80% ribavirin for more than 80% expectedduration of therapy. d Represents the subgroup receiving 80% peginterferon -2b and/or ribavirin for more than 80% expected duration of therapy. e Alanine aminotransferase expressedas a ratio to the upper limit of normal (ULN). levels of adherence for all patients and patients with HCV-1who received interferon -2b and ribavirin, as well as for all patients and patients with HCV-1who received interferon alone. The adherence-response regression curves for the combination therapy groups and the monotherapy groups are superimposed. The difference between the 2 regression curves was used to estimate the effect of combination therapy at 100% adherence. The resulting estimate, marked with an X on the plot, is given in Table 2. Figure 3 shows the analogous analyses for all patients and HCV-1patients who received peginterferon -2b plus ribavirin or peginterferon alone, and Figure 4 shows the analysis for patients who received the weight-based dosing regimen of peginterferon -2b plus ribavirin or peginterferon -2b alone. Estimates based on statistical modeling of the fulldose sustained-response rates for interferon -2b plus ribavirin, peginterferon -2b plus ribavirin 800 mg/day, and peginterferon -2b plus weight-based ribavirin are 50%, 62%, and 71%, respectively. These are close to the respective rates observed for the 80/80/80 adherent subgroup at 52%, 63%, and 72%. The same agreement in sustained virologic response rates as estimated by the 2 approaches occurs when we restrict the population to patients with HCV-1. The close agreement between the 2 approaches supports the conclusion that patients who are at least 80% adherent will derive maximum benefit from treatment. In addition, we observed progressively higher sustained response rates with increasing levels of adherence, irrespective of the treatment regimen. However, the adherence-response regression curves cannot be interpreted as dose-response curves and do not convey any relationship between dose taken and efficacy response outside of the clinically controlled environment. Thus the model is unable to estimate SVR rates at different adherence levels. The effect of early and late dose reductions was analyzed in the patient groups receiving interferon or peginterferon and ribavirin (Figure 5). Whereas most patients who reduced their dose within the first 12 weeks of therapy maintained a reduced dose during the remainder of therapy, few if any patients who were not adherent during the first 12 weeks became adherent thereafter. A trend was observed between early dose reductions and impaired sustained response rates; this trend was less apparent for those who required dose reductions late in the course of therapy. But the results of this analysis were not statistically significant. We attempted to analyze the effect of dose reduction of each drug on the sustained response rate; however, 80% of the patients who received 80% of either interferon -2b or peginterferon -2b also received 80% of the doses of ribavirin. Thus

6 1066 McHUTCHISON ET AL. GASTROENTEROLOGY Vol. 123, No. 4 Table 2. Sustained Virologic Response Rates According to Adherence Regimen ITT a A 80/80/80 b B 80/ 80/ EstimatedSR; 80 c C Excluded d full adherence e A vs. B f B vs. C f IFN -2b ribavirin g All Patients 440/1010 (44%) 326/631 (52%) 75/161 (47%) % Genotype 1 (48 weeks) 208/668 (31%) 165/423 (39%) 34/108 (31%) % Genotype 2 or 3 (24 weeks) 105/157 (67%) 91/127 (72%) 11/14 (79%) Genotype 2 or 3 (48 weeks) 102/157 (65%) 67/96 (70%) 15/19 (79%) PegIFN -2b 1.5 g/kg ribavirin h All patients (48 weeks) 274/511 (54%) 191/305 (63%) 61/118 (52%) 88 62% Genotype 1 145/348 (42%) 105/206 (51%) 27/79 (34%) 63 50% Genotype 2 or 3 121/147 (82%) 79/88 (90%) 34/38 (89%) PegIFN -2b 1.5 g/kg ribavirin h 10.6 mg/kg All Patients (48 weeks) 114/188 (61%) 76/106 (72%) 29/51 (57%) 31 71% Genotype 1 58/122 (48%) 42/67 (63%) 11/32 (34%) 23 61% Genotype 2 or 3 51/58 (88%) 30/32 (94%) 18/19 (95%) PegIFN -2b 1.5 g/kg i All patients (48 weeks) 71/304 (23%) 63/236 (27%) 6/23 (26%) Genotype 1 31/223 (14%) 30/175 (17%) 1/14 (7%) Genotype 2 or 3 36/73 (49%) 30/55 (54%) 4/7 (57%) Interferon -2b i All patients 37/303 (12%) 37/242 (15%) 0/5 (0%) Genotype 1 14/217 (6%) 14/170 (8%) 0/3 (0%) Genotype 2 or 3 23/81 23/67 0/ (28%) (34%) (0%) a Denotes the sustainedvirologic response rate by intent to treat or primary analysis. b Indicates the subgroup that received 80% interferon or peginterferon plus 80% ribavirin for more than 80% of the expectedduration of therapy. c Indicates the less adherent group subgroup who received 80% interferon or peginterferon and/or 80% ribavirin but for 80% expected duration of therapy. d Denotes the number of patients excludedfrom these analyses because they receivedtherapy for less than 80% of the expectedtreatment duration. Thus, patients who received 1 or 2 days of treatment or shortened courses of therapy were excluded. e Denotes the estimatedsustainedresponse rate by the statistical model as describedfor complete adherence in each group of patients receiving each type of therapy depicted. f Comparisons between groups shown as the calculated P value by 2 analyses. g Data from the McHutchison 3 andpoynard 4 interferon andribavirin combination therapy trials. h Data from the Manns 7 peginterferon -2b andribavirin combination therapy trial. i Data from the Lindsay 10 peginterferon -2b monotherapy trial. we were unable to distinguish which drug was more important in terms of the impact of adherence on sustained response. Discussion This retrospective analysis indicates that adherence to therapy with interferon -2b or peginterferon -2b plus ribavirin for patients with chronic hepatitis C is important and enhances sustained response rates. This observation was apparent only for HCV-1 infected patients, those most difficult to treat. Moreover, the fact that results obtained with 2 different approaches subgroup analysis of observed data and modeling produced similar results strengthens the premise that adherent patients are more likely to achieve a sustained virologic response. This study does not replace the primary ITT analysis, but does address the important issue of adherence to prescribed therapy. Most patients in these clinical registration trials managed to achieve the goals of 80% adherence to their medication dose and duration of therapy. Because of these patients motivation and their management in controlled trials at tertiary referral centers, however, our findings may not be representative or applicable to the larger universe of patients managed in clinical practice. Thus patients treated in community practices or in centers with limited expertise in the management of chronic hepatitis C may have lower response rates than those observed in this study. Nevertheless, it seems reasonable that the concept of adherence to therapy should also apply to patients treated outside the realm of clinical trials. Further prospective studies are needed to address this issue. Although we attempted to evaluate the effect of adherence for each drug individually and of early adherence vs. late adherence to therapy, these analyses were limited

7 October 2002 ADHERENCE TO THERAPY AND HEPATITIS C 1067 Figure 2. (A) The observedsustainedvirologic response rates at different levels of adherence for all patients who received interferon -2b plus ribavirin (solid line) andfor all patients who receivedinterferon alone (dashed line). These corresponding observed response rates (as a percentage) are listedbelow the plot under the appropriate adherence levels. The ITT response rate for patients who received interferon andribavirin is 44% (440 of 1010) andthat for patients who receivedinterferon alone is 16% (82 of 503). The adherence-response regression curves for the 2 treatment groups are superimposed. The estimated full-adherence sustained response rate for combination therapy is 50% (denoted by X ). (B) The relationship between observedadherence andsustainedvirologic response rates for the subset of HCV-1 patients. The ITT response rate for HCV-1 patients who receivedinterferon andribavirin is 31% (208 of 668), andthat for patients who receivedinterferon alone is 9% (30 of 330). The estimatedfull-adherence effect for combination therapy for HCV-1 patients is 34% (denoted by X ). by the fact that most patients who were adherent to 1 drug were adherent to both drugs. Also, very few patients who were nonadherent during the first 12 weeks of therapy became adherent later, preventing us from clarifying the importance of the timing of adherence. Thus this retrospective analysis does not allow us to define the effect of early versus late adherence or the contribution of adherence to each drug individually. Although our subgroup analysis showed only a trend toward improving sustained response in adherent patients treated with interferon -2b and ribavirin, the results of statistical modeling showed similar results irrespective of the treatment regimen. Adherent patients receiving interferon -2b and ribavirin were more likely to be males and weighed more, whereas those who were less adherent tended to be older and more frequently had bridging fibrosis or cirrhosis. Reasons for these baseline differences in gender and body weight between the adherent and nonadherent groups are unknown and require further prospective evaluation. HCV-2 or -3 infected patients demonstrated excellent virologic response rates to combination therapy according to the ITT analysis, and the detection of significant incremental virologic response rates with adherence would require larger numbers than were evaluated in this study. Figure 3. (A) The observedsustainedresponse rates at different levels of adherence for all patients who received peginterferon -2b plus ribavirin (solid line) andall patients who receivedpeginterferon alone (dashed line). The adherence-response regression curves are superimposed. The ITT response rate is 54% (274 of 511) for peginterferon plus ribavirin and23% (71 of 304) for peginterferon alone. The estimatedfull-adherence response rate for combination therapy is 62% (denoted by X ). (B) The relationship between observed adherence andsustainedvirologic response rates for the subset of patients infectedwith HCV-1. The ITT response rate is 42% (145 of 348) for combination therapy and14% (31 of 223) for monotherapy. The estimatedfull-adherence sustainedresponse rate for combination therapy in patients with HCV-1 is 50% (denoted by X ). Figure 4. (A) The observedsustainedresponse rates at different levels of adherence for patients who receivedpeginterferon andribavirin at a dose of at least 10.6 mg/kg and corresponding patients of the same weight group who receivedpeginterferon alone. The adherence-response regression curves are superimposed. The ITT response rate is 61% (114 of 188) for the weight-basedsubgroup that receivedpeginterferon plus ribavirin and25% (30 of 118) for the subgroup that receivedpeginterferon alone. The estimatedfull-adherence response rate for weight-basedcombination therapy is 71% (denoted by X ). (B) The relationship between observedadherence andsustainedvirologic response rates for the subset of patients also infectedwith HCV-1. The ITT response rate is 48% (58 of 122) for weight-basedcombination therapy and14% (12 of 83) for the monotherapy. The estimatedfull-adherence sustainedresponse rate for weight-basedcombination therapy in patients with HCV-1 is 61% (denoted by X ).

8 1068 McHUTCHISON ET AL. GASTROENTEROLOGY Vol. 123, No. 4 Figure 5. Contingency tables indicating the effect of drug adherence during the first 12 weeks of therapy compared to that during the remainder of treatment andthe impact on sustainedviral response rates. (A) Observeddata for the group treatedwith interferon -2b plus ribavirin; (B) depicts data for patients treated with peginterferon -2b plus ribavirin. Most patients were adherent for the duration of treatment; a minority of patients who were not adherent in the first 12 weeks were adherent thereafter. Adherence to therapy for the first 12 weeks improves response rates comparedto nonadherence throughout treatment. In HIV infection and other diseases, adherence to therapy is related to the number of medications taken per day, dietary restrictions required for these medications, dosing frequency, pill size, and the drug combinations used. 16,17 In addition, low literacy level, beliefs regarding therapy (whether erroneous or otherwise), lack of support, inconvenient appointments, and lack of transportation have all been associated with a lack of adherence to therapy in HIV-infected patients. 9,18 In HIV infection, the most common reasons for lack of adherence or skipping doses are simply forgetting, being too busy, side effects, and feeling ill, 9 all of which would seem to apply to patients with chronic hepatitis C undergoing therapy. Demographic variables, such as gender, age, race, socioeconomic status, and history of prior substance abuse, generally do not predict poor adherence to HIV therapy. 19 Such variables associated with nonadherence notwithstanding, physicians are able to accurately predict adherence in only about 50% of patients. 20 Thus methods to monitor and improve adherence may help improve the likelihood of successful outcome during and after treatment of any disease process. Although no gold standard exists for monitoring adherence in clinical trials or in practice, theoretically, directly observed therapy 21 (as previously used for tuberculosis) may be impracticable in hepatitis C treatment regimens, but it could be made available to a limited number of patients considered to be at high risk for nonadherence. Although currently impractical in hepatitis C treatment regimens, directly observed therapy could be facilitated by once-weekly peginterferon dosing regimens. Patient selfreporting, 22 as in our studies, is less accurate than other potential methods but is a common and convenient way of aiding adherence. A medication diary is also an important behavioral tool that may help reinforce adherence. Electronic monitoring devices, such as medication event monitoring systems, may also increase adherence, and future studies of hepatitis C therapy involving these microprocessor recorders could provide a better assessment of the effect of adherence. In HIV therapy, encouraging telephone calls and feedback about viral loads have proven to improve adherence, but evidence to support their approach in hepatitis C therapy is lacking. Still, these approaches merit consideration. A number of other factors may impair treatment adherence in hepatitis C, including language barriers, negative staff attitudes toward injecting drug users, and social circumstances, such as lack of family support. Strategies to improve adherence include encouraging reasonable adherence instead of demanding full adherence, involving family members or significant others in the treatment process, and anticipating presumptively and responding to adverse social situations that could reduce adherence. Although patients themselves are the primary determinants of adherence, the importance of the role of health care providers should not be overlooked. This study also suggests an improved sustained virologic response with increasing adherence. It may be possible to use a similar adherence-efficacy model in future clinical trials to gauge the impact of patient adherence on therapeutic outcome, and potentially to determine ideal drug dosages for those most compliant to therapy. Our study was not designed to evaluate adherence-response rates with varying treatment duration. As such, the current recommendations for duration of therapy should remain the same, even for compliant patients. The findings of this study indicate that adherence with prescribed medication regimens does improve sustained response rates in patients with chronic hepatitis C infection. Further research must now be conducted in prospective trials to determine factors associated with adherence or lack thereof and strategies for improving adherence, in an attempt to enhance sustained response rates. Theoretically, patient education about side effects, necessary lifestyle changes during therapy, treatment of depression, support groups, telephone and frequent clinic follow-up visits, printed materials, pill boxes, reminders, and self-monitoring devices, as well as simplification of treatment regimens, all have the potential to improve adherence. Certainly, the introduction of peginterferon, by reducing the number and frequency of injections, has also simplified the treatment regimen, which may facilitate adherence to therapy. Although more work in improving adherence is necessary, our results suggest that adherence will enhance the likelihood of achieving a virologic response during

9 October 2002 ADHERENCE TO THERAPY AND HEPATITIS C 1069 the first 3 6 months of therapy, and that maintaining adherence in patients who have responded to treatment by week will lead to an increase in sustained virologic response rates. References 1. Alter M, Kruszon-Moran D, Nainan O, McQuillan GM, Gao F, Moyer LA, Kaslow RA, Margolis HS. The prevalence of hepatitis C virus in the UnitedStates, 1988 through N Engl J Med1999; 341: Detre K, Belle S, Lombardero M. Liver transplantation for chronic viral hepatitis. Viral Hepat Rev 1996;2: McHutchison JG, Gordon SC, Schiff ER, Shiffman ML, Lee WL, Rustgi VK, Goodman ZD, Ling M-H, Cort S, Albrecht JK. Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. Hepatitis Interventional Therapy Group. N Engl J Med1998;339: Poynard T, Marcellin P, Lee S, Niederau C, Minuk GS, Ideo G, Bain V, Heathcote J, Zeuzem S, Trepo C, Albrecht J for the International Hepatitis Interventional Therapy Group. Randomised trial of interferon alfa-2b plus ribavirin for 48 weeks or 24 weeks versus interferon alfa-2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. Lancet 1998;352: Marcellin P, Boyer N, Gervais A, Martinot M, Pouteau M, Castelnau C, Kilani A, Areias J, Auperin A, Benhamou JP, Degott C, Erlinger S. Long-term histologic improvement andloss of detectable intrahepatic HCV RNA in patients with chronic hepatitis C andsustainedresponse to interferon-alpha therapy. Ann Intern Med1997;127: Lau DTY, Kleiner DE, Ghany MG, Park Y, SchmidP, Hoofnagle JH. 10-year follow-up after interferon-alpha therapy for chronic hepatitis C. Hepatology 1998;28: Manns MP, McHutchison JG, Gordon SC, Rustgi VK, Shiffman M, Reindollar R, Goodman ZD, Koury K, Ling M-H, Albrecht JK for the International Hepatitis Interventional Therapy Group. Peginterferon alfa-2b in combination with ribavirin comparedto interferon alfa2b plus ribavirin for initial treatment of chronic hepatitis C: results of a randomized trial. Lancet 2001;358: Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C, Wagener MM, Singh N. Adherence to protease inhibitor therapy andoutcomes in patients with HIV infection. Ann Intern Med 2000;133: Chesney MA, Ickovics J, Hecht FM, Sikipa G, Rabkin J. Adherence: a necessity for successful HIV combination therapy. AIDS 1999;13(suppl 1):S271 S Lindsay KL, Trepo C, Heintges T, Shiffman ML, Gordon SC, Hoefs JC, Schiff ER, Goodman ZD, Laughlin M, Yao R, Albrecht JK for the Hepatitis Interventional Therapy Group. A randomized, double-blindtrial comparing pegylatedinterferon alfa-2b to interferon alfa-2b as initial treatment for chronic hepatitis C. Hepatology 2001;34: Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N, Kiernan TW, Wollman J. Formulation andapplication of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis. Hepatology 1981;5: Tong MJ, Hwang SJ, Lefkowitz M, Lee SD, Co RL, ConradA, SchmidP, Lo KJ. Correlation of serum HCV RNA andalanine aminotransferase levels in chronic hepatitis C patients during treatment with ribavirin. J Gastroenterol Hepatol 1994;9: Stuyver L, Rossau R, Wyseur A, Duhamel M, Vanderborght B, Van Heuverswyn H, Maertens G. Typing of hepatitis C virus isolates andcharacterization of new subtypes using a line probe assay. J Gen Virol 1993;74: Efron B, Feldman D. Compliance as an explanatory variable in clinical trials (with discussion). J Am Stat Assoc 1991;86: Mak C, Heft S, Amkraut H, Ling M. Compliance as an explanatory variable in hepatitis C. Drug Inform J 2001;35: Bartlett J DR, DeMasi R,Quinn J, Moxham C, Rousseau F. Correlation between antiretroviral pill burden and durability of virologic response: a systematic overview. In Abstracts of the XIII International AIDS Conference, July 9 14, 2000, Durban, South Africa, 2000; abstract THPeB Eisen SA, Miller DK, Woodward RS, Spitznagel E, Przybeck TR. The effect of prescribed daily dose frequency on patient medication compliance. Arch Intern Med1990;150: Kalichman SC, Ramachandran B, Catz S. Adherence to combination antiretroviral therapies in HIV patients of low health literacy. J Gen Intern Med1999;14: Klaus BD, Grodesky MJ. Assessing and enhancing compliance with retroviral therapy. Nurse Pract 1997;22: Eldred LJ, Wu AW, Chaisson RE, Moore RD. Adherence to antiretroviral andpneumocystis prophylaxis in HIV disease. J Acquir Immune Defic Syndr Hum Retrovirol 1998;18: Friedland GH, Williams A. Attaining higher goals in HIV treatment: the central importance of adherence. AIDS 1999;13(suppl 1): S61 S Bangsberg DR, Hecht FM, Charlebois ED, Zolopa AR, Holodniy M, Sheiner L, Bamberger JD, Chesney MA, Moss A. Adherence to protease inhibitors, HIV-1 viral load, and development of drug resistance in an indigent population. AIDS 2000;14: Received December 28, Accepted July 11, Address requests for reprints to: John G. McHutchison, M.D., Duke Clinical Research Institute, Duke University Medical Center, P.O. Box 17969, Durham, North Carolina mchutool@ mc.duke.edu; fax: (919) Supported in part by research grants from the Schering-Plough Research Institute, Kenilworth, New Jersey, and clinical research center grants from Massachusetts General Hospital (MO1-RR01066), the Scripps Clinic (MO1-RR00833), and Los Angeles County University of Southern California Medical Center (M01-RR00043). Dr. McHutchison is an advisor and consultant and has received research support from Schering-Plough. Dr. Manns is an ad hoc consultant to and a recipient of research grants from Schering-Plough. Dr. Poynard works on the Principal Investigator and Speaker Bureau for Schering-Plough. Dr. Lindsay is a speaker for Schering-Plough and has received research support from Schering-Plough, Glaxo Wellcome, Hoffman-Rodel, Scripps Liver Research Consortium, and Triangle Pharmaceuticals. Dr. Dienstag has received research support and has served on an adjudication board for Schering-Plough. Drs. Mak and Albrecht are employees of Schering-Plough and own stock in the corporation. The authors thank Elvia Nunez of the Scripps Clinic for preparing this manuscript. The authors also thank all participating investigators in these randomized controlled trials. Their individual names and sites appear in the appendices attached to each of the separate publications.